Design and Verification of Specialised Security Goals for Protocol Families

Doctoral thesis (2020)

Communication Protocols form a fundamental backbone of our modern information networks. These protocols provide a framework to describe how agents - Computers, Smartphones, RFID Tags and more - should ... [more ▼]

Communication Protocols form a fundamental backbone of our modern information networks. These protocols provide a framework to describe how agents - Computers, Smartphones, RFID Tags and more - should structure their communication. As a result, the security of these protocols is implicitly trusted to protect our personal data. In 1997, Lowe presented ‘A Hierarchy of Authentication Specifications’, formalising a set of security requirements that might be expected of communication protocols. The value of these requirements is that they can be formally tested and verified against a protocol specification. This allows a user to have confidence that their communications are protected in ways that are uniformly defined and universally agreed upon. Since that time, the range of objectives and applications of real-world protocols has grown. Novel requirements - such as checking the physical distance between participants, or evolving trust assumptions of intermediate nodes on the network - mean that new attack vectors are found on a frequent basis. The challenge, then, is to define security goals which will guarantee security, even when the nature of these attacks is not known. In this thesis, a methodology for the design of security goals is created. It is used to define a collection of specialised security goals for protocols in multiple different families, by considering tailor-made models for these specific scenarios. For complex requirements, theorems are proved that simplify analysis, allowing the verification of security goals to be efficiently modelled in automated prover tools. [less ▲]

Breaking Unlinkability of the ICAO 9303 Standard for e-Passports using Bisimilarity

in Breaking Unlinkability of the ICAO 9303 Standard for e-Passports using Bisimilarity (2019, September 23)

We clear up confusion surrounding privacy claims about the ICAO 9303 standard for e-passports. The ICAO 9303 standard includes a Basic Access Control (BAC) protocol that should protect the user from being ... [more ▼]

We clear up confusion surrounding privacy claims about the ICAO 9303 standard for e-passports. The ICAO 9303 standard includes a Basic Access Control (BAC) protocol that should protect the user from being traced from one session to another. While it is well known that there are attacks on BAC, allowing an attacker to link multiple uses of the same passport, due to differences in implementation; there still remains confusion about whether there is an attack on unlinkability directly on the BAC protocol as specified in the ICAO 9303 standard. This paper clarifies the nature of the debate, and sources of potential confusion. We demonstrate that the original privacy claims made are flawed, by uncovering attacks on a strong formulation of unlinkability. We explain why the use of the bisimilarity equivalence technique is essential for uncovering our attacks. We also clarify what assumptions lead to proofs of formulations of unlinkability using weaker notions of equivalence. Furthermore, we propose a fix for BAC within the scope of the standard, and prove that it is correct, again using a state-of-the-art approach to bisimilarity. [less ▲]

Distance-Bounding Protocols: Verification without Time and Location

in Proceedings of IEEE Symposium on Security and Privacy (SP), San Francisco 21-23 May 2018 (2018)

Distance-bounding protocols are cryptographic protocols that securely establish an upper bound on the physi- cal distance between the participants. Existing symbolic verification frameworks for distance ... [more ▼]

Distance-bounding protocols are cryptographic protocols that securely establish an upper bound on the physi- cal distance between the participants. Existing symbolic verification frameworks for distance-bounding protocols consider timestamps and the location of agents. In this work we introduce a causality-based characterization of secure distance-bounding that discards the notions of time and location. This allows us to verify the correct- ness of distance-bounding protocols with standard pro- tocol verification tools. That is to say, we provide the first fully automated verification framework for distance- bounding protocols. By using our framework, we con- firmed known vulnerabilities in a number of protocols and discovered unreported attacks against two recently published protocols. [less ▲]